Viscoelastic Properties of Cationic Starch Adsorbed on Quartz Studied

added NaCl 0, 1, 100, and 500 mM) on silica were studied with a quartz crystal microbalance with dissipation. (QCM-D). Three different starches were i...
0 downloads 5 Views 554KB Size
10900

Langmuir 2004, 20, 10900-10909

Viscoelastic Properties of Cationic Starch Adsorbed on Quartz Studied by QCM-D Tekla Tammelin,*,† Juha Merta,†,§ Leena-Sisko Johansson,‡ and Per Stenius† Laboratory of Forest Products Chemistry, Helsinki University of Technology, Espoo, Finland, and HUT Center for Chemical Analysis, Helsinki University of Technology, Espoo, Finland Received May 18, 2004. In Final Form: August 23, 2004 The adsorption and viscoelastic properties of layers of a cationic polyelectrolyte (cationic starch, CS, with 2-hydroxy-3-trimethylammoniumchloride as the substituent) adsorbed from aqueous solutions (pH 7.5, added NaCl 0, 1, 100, and 500 mM) on silica were studied with a quartz crystal microbalance with dissipation (QCM-D). Three different starches were investigated (weight-average molecular weights Mw ≈ 8.7 × 105 and 4.5 × 105 with degree of substitution DS ) 0.75 and Mw ≈ 8.8 × 105 with DS ) 0.2). At low ionic strength, the adsorbed layers are thin and rigid and the amount adsorbed can be calculated using the Sauerbrey equation. When the ionic strength is increased, significant changes take place in the amount of adsorbed CS and the viscoelasticity of the adsorbed layer. These changes were analyzed assuming that the layer can be described as a Voigt element on a rigid surface in contact with purely viscous solvent. It was found that CS with low charge density forms a thicker and more mobile layer with higher viscosity and elasticity than CS with high charge density. The polymers adsorbed on the silica even when the ionic strength was so high that electrostatic interactions were effectively screened. At this high ionic strength, it was possible to study the effect of molecular weight and molecular weight distribution of the CS on the properties of the adsorbed film. Increasing the molecular weight of CS resulted in a larger hydrodynamic thickness. CS with a narrow molecular weight distribution formed a more compact and rigid layer than broadly distributed CS, presumably due to the better packing of the molecules.

Introduction Cationically modified starch (CS) has widespread use as a papermaking chemical for improving retention and the dry strength of paper. Hence, it is of interest to gain a proper understanding of how adsorption and the properties of the adsorbed layers of CS at the solid/liquid interface are affected by parameters such as ionic strength, surface charge density, and the molecular weight and degree of substitution (charge density) of the CS. This was the goal of the study described in this paper. While there is an extensive literature on the theory of adsorption of polyelectrolytes from solution on oppositely charged surfaces1 as well as a large number of reports on experimental investigations of polycation adsorption, including cationic starch,2-4 much less is known about the actual physical properties of the adsorbed layers. Previously, we have shown that cationic starch apparently retains the helical conformation of native starch in aqueous solution.19 Thus, knowledge about the structure of adsorbed layers of CS is of interest not only due to its practical importance but also because the conformational changes taking place when molecules of this type are adsorbed are not well-known. As the main method of investigation, we have used the QCM-D instrument (quartz crystal microbalance with * Corresponding author. Tel: +358-9-4514238. Fax: +358-94514259. E-mail: [email protected]. † Laboratory of Forest Products Chemistry. ‡ HUT Center for Chemical Analysis. § Present address: KCL Science and Consulting, Espoo, Finland. (1) Fleer, G. J.; Stuart, M. A. C.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at interfaces; Chapman & Hall, University Press: Cambridge, 1993. (2) van de Steeg, H. G. M.; Stuart, M. A. C.; de Keizer, A.; Bijsterbosch, B. H. Langmuir 1992, 8, 2538-2546. (3) O ¨ dberg, L.; Sandberg, S. Langmuir 1995, 11, 2621-2625. (4) Bremmell, K. E.; Jameson, G. J.; Biggs, S. Colloids Surf., A 1998, 139, 199-211.

dissipation5). The QCM has traditionally been used as a mass sensor of adsorption of, for example, proteins6 and synthetic polyelectrolytes from solution.7 The QCM-D instrument offers new possibilities of in situ investigation of not only adsorption kinetics and adsorbed mass but also the time dependence of viscous and elastic properties of adsorbed polymer layers at the solid/liquid interface.8,9 X-ray photoelectron spectroscopy, XPS, was used as an additional method for quantification of the adsorbed mass. Experimental Section Materials. Cationic Starch. CS was synthesized from potato starch at the laboratories of Raisio Chemicals Oy, Raisio, Finland. Before cationization, the starch was partially depolymerized by reaction with sodium hypochlorite in a slurry of the starch at controlled pH and temperature.10 The acidic reaction product was neutralized with dilute NaOH. Excess hypochlorite was decomposed by addition of sodium bisulfite. The starch was then reacted with 1-chloro-2-hydroxy-3-trimethylammonium-propyl chloride. The resulting product was purified by filtration in a tangential flow ultrafiltration system (Filtron Technology Corp., Northboro, MA; Minisette), using a membrane with a cutoff of 8000. Three samples were prepared: one with low degree of substitution (DS) and high molecular weight (Mw), one with high DS and high Mw, and one with high DS and low Mw. Some properties of the CS samples are given in Table 1. Molecular weight distributions determined by size exclusion chromatography (SEC) are shown in Figure 1. (5) Rodahl, M.; Kasemo, B. Sens. Actuators, A 1996, 54, 448-456. (6) Tanaka, M.; Mochizuki, A.; Motomura, T.; Shimura, K.; Onishi, M.; Okahata, Y. Colloids Surf., A 2001, 193, 145-152. (7) Baba, A.; Kaneko, F.; Advincula, R. C. Colloids Surf., A 2000, 173, 39-49. (8) Rodahl, M.; Ho¨o¨k, F.; Fredriksson, C.; Keller, C. A.; Krozer, A.; Brzezinski, P.; Voinova, M.; Kasemo, B. Faraday Discuss. 1997, 107, 229-246. (9) Plunkett, M. A.; Claesson, P. M.; Rutland, M. W. Langmuir 2002, 18, 1274-1280. (10) Rutenberg, M. W.; Solarek, D. Starch derivatives: Production and uses. In Starch Chemistry and Technology; Whistler, R. L., Miller, J. N., Paschall, E. F., Eds.; Academic Press: Orlando, FL, 1984; pp 311-388.

10.1021/la0487693 CCC: $27.50 © 2004 American Chemical Society Published on Web 11/06/2004

Viscoelastic Properties of Cationic Starch

Langmuir, Vol. 20, No. 25, 2004 10901

Figure 1. The molecular weight distribution of cationic starches. Key: ([) high Mw, low DS, narrow Mw distribution, CSLH; (‚‚‚‚) high Mw, high DS, broad Mw distribution, CSHH; (- - -) low Mw, high DS, narrow Mw distribution, CSHL. Table 1. Properties of Cationic Starch Samples CSHLa pHd

10.7 weight-average molecular wte 8.8 × 105 number-average molecular wte 4.3 × 105 degree of substitution (DS)d 0.20

CSHHb

CSLHc

10.8 8.7 × 105 3.5 × 105 0.75

10.8 4.5 × 105 2.0 × 105 0.75

a High molecular weight, low DS. b High molecular weight, high DS. c Low molecular weight, high DS. d As reported by the manufacturer. e By SEC.

Solutions. Solutions of CS in water were prepared by heating a mixture of CS and water in an autoclave for 10 min at 120 °C. All solutions were prepared at least 24 h before measurements. Starch concentrations used in the QCM-D experiments were 1, 10, and 100 ppm. The ionic strength of the solutions without added salt was 0.1-0.3 mM. In studies of the effect of salt on the adsorption, the concentrations of added simple electrolyte were 1, 100, and 500 mM NaCl. The pH of the starch solution was 7.5. Other Chemicals. The water was distilled. All other chemicals were analytical grade and were used without further purification. QCM-D Crystals. The QCM-D crystals used were so-called AT-cut quartz crystals supplied by Q-sense AB, with thickness 0.3 mm, fundamental frequency f0 ≈ 5 MHz, and sensitivity constant C ) 0.177 mg m-2 Hz-1. The crystals were coated with silica by means of vapor deposition. Thus silica was always the adsorbent surface. Methods. Quartz Crystal Microbalance. Adsorption and the properties of the adsorbed layer were studied using a quartz crystal microbalance with dissipation, with a QCM-D instrument from Q-sense Ab, Gothenburg, Sweden.11 The principle of this instrument is the following. Without adsorbate, the crystal, when immersed in aqueous solution, oscillates at a resonant frequency f0. When material adsorbs on the crystal, the resonance frequency is lowered to f. The instrument measures the shift in the frequency of the fundamental resonance and several overtones. If the adsorbed material is evenly distributed, rigidly attached, and small compared to the mass of the crystal, the frequency shift ∆f ) f - f0 is related to the adsorbed mass per unit surface, ∆m, by the Sauerbrey equation:12

Figure 2. Model used for interpretation of viscoelastic properties (ref 12). (a) A viscoelastic thin film of density Ff and thickness hf is located between the elastic quartz crystal of density Fq and thickness hq and a viscous solution of viscosity ηl and density Fl. (b) The film is subjected to an oscillating shear stress σ and behaves like a Voigt element with shear viscosity ηf and shear modulus µf.

(1)

When the source driving the oscillating crystal is cut off, the amplitude of the oscillation decays due to frictional losses (dissipation of energy) in the crystal, in the adsorbed layer, and in the surrounding solution. The decay rate depends on the viscoelastic properties of these materials. Dissipation is char-

where n is the overtone number (in the present case n ) 1, 3, or 5) and C is a constant that describes the sensitivity of the device to changes in mass.

(11) Rodahl, K.; Ho¨o¨k, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. Rev. Sci. Instrum. 1995, 66, 3924. (12) Ho¨o¨k, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 7290.

∆m ) -

C∆f n

10902

Langmuir, Vol. 20, No. 25, 2004

Tammelin et al.

Figure 3. An example of an XPS survey recorded for a layer of low molecular weight and high DS (CSLH) cationic starch adsorbed on SiO2 from a 100 ppm solution in 100 mM NaCl. The excerpt shows the three measurements of the nitrogen N 1s in the same sample. acterized by the dissipation factor Dd, which is defined by

Ediss Dd ) 2πEstor

(2)

where Ediss is the dissipated energy during one oscillation cycle and Estor is the total energy stored in the oscillation. With the QCM-D instrument, the change in the dissipation factor, ∆D ) D - D0, is measured, where D0 is the dissipation factor of the pure quartz crystal immersed in the solvent and D is the dissipation factor when material has been adsorbed. Interpretation of Viscoelastic Properties. Using appropriate models, ∆f and ∆D can be interpreted in terms of adsorbed mass and structural changes in the adsorbed layer. The complex shear modulus of the film is defined by

G ) G′ + iG′′ ) µf + 2πifηf ) µf(1 + 2πifτf)

(3)

where µf is the elastic shear (storage) modulus, ηf is the shear viscosity (loss modulus), f is the oscillation frequency, and τf is the characteristic relaxation time of the film. In this work, the interpretation of viscoelastic properties of the adsorbed layer film was based on the model presented by Voinova et al.13 In this model (Figure 2), the adsorbed film is represented by a single Voigt element, the quartz crystal is assumed to be purely elastic, and the surrounding solution is assumed to be purely viscous and Newtonian. Further, it is assumed that the thickness hf and the density Ff of the film are uniform, that the viscoelastic properties are frequency independent, and that there is no slip between the adsorbed layer and the crystal during shearing. Then ∆f and ∆D are given by

(

∆f ) Im

)

(4)

( )

(5)

β 2πFqhq

and

∆D ) -Re

β πFqhq

where Fq and hq are the density and thickness of the quartz plate, and

2πfηf - iµf 1 - R exp(2ξ1hf) β ) ξ1 2πf 1 + R exp(2ξ1hf)

(6)

ξ1 2πfηf - iµf +1 ξ2 2πfηl , R) ξ1 2πfηf - iµf -1 ξ2 2πfηl

ξ1 )

x

-

(2πf)2Ff , µf + i2πfηf

ξ2 )

x

2πfFl (7) ηl

i

where the indexes q, f, and l refer to the crystal, film, and bulk liquid, respectively. Results from measurements at several overtones were fitted to this model using the program Q-Tools from Q-sense Ab. Size Exclusion Chromatography. Molecular weight distributions were analyzed using size exclusion chromatography. The eluent was a NaNO3-NaH2PO4 buffer with a flow rate of 1 mL/ min. Molecular weight distributions were calculated on the basis of the uniform pullullan standard. The columns used were Separon HEMA-BIO 10, Separon HEMA-BIO 300, and Separon HEMA-BIO 1000 HPSEC connected in series. X-ray Photoelectron Spectroscopy. The ex situ XPS measurements were performed with a Kratos Analytical AXIS 165 electron spectrometer using a monochromated Al KR X-ray source at the HUT Center for Chemical Analysis. The experiments were carried out on freshly prepared, dried samples according to the standardized procedure specifically developed at HUT for cellulosic materials.14,15 In this XPS study, two markers for the CS were used: the surface content of nitrogen and the relative abundance of C-O, that is, carbon bonded with only one oxygen atom. Nitrogen was determined from the N 1s trace analysis, measured with lowresolution settings identical to those used in the survey spectra (see Figure 3). C-O was determined from curve-fitted highresolution C 1s regions (using a fit with four symmetric Gaussian peaks regularly applied for celluloses; see Figure 4). There are some problems associated with both of the XPS markers used. In the case of nitrogen, the amounts were quite small for quantitative analysis, despite the extended acquisition times. On the other hand, C-O bonds have been shown to yield reproducible data on carbohydrates for celluloses, but the total carbon signal may be biased toward C-C carbon due to airborne contamination present on SiO2. Hence, to improve the statistics (13) Voinova, M.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391-396. (14) Johansson, L.-S. Microchim. Acta 2002, 138, 217. (15) Koljonen, K.; O ¨ sterberg, M.; Johansson, L.-S.; Stenius, P. Colloids Surf., A 2003, 228, 147.

Viscoelastic Properties of Cationic Starch

Langmuir, Vol. 20, No. 25, 2004 10903

Figure 4. High-resolution regions of O 1s and C 1s. Table 2. Change in Frequency and Dissipation at Equilibrium and the Adsorbed Mass Calculated from Equation 1a CSHL Mw ) 8.8 × 105, DS ) 0.2 NaCl (mM)

∆f (Hz)

0 -10.5 1 -24.2 100 -100

CSLH Mw ) 4.5 × 105, DS ) 0.75

∆m ∆m ∆f ∆D × 10-6 (mg m-2) (Hz) ∆D × 10-6 (mg m-2) 0.14 0.42 5.94

0.62 1.43

-6.0 -6.4 -73.7

0.04 0.23 4.21

0.35 0.38

a Adsorption from 100 ppm CS in different electrolytes; f ) 5 0 MHz, n ) 3.

and also to average out the heterogeneities of the samples, both survey scans and the regional spectra were recorded at three different spots on each sample. All spectra were collected at an electron takeoff angle of 90° from sample areas less than 1 mm in diameter and using only 100 W irradiation. The total exposure times were kept below 20 min per spot, since long exposures to X-rays are known to easily degrade organic specimens.15 No sample degradation was detected under these circumstances.16

Results and Preliminary Interpretation The Effect of Degree of Substitution and Electrolyte Concentration. QCM-D Measurements. After replacement of water in the QCM cell with a solution of cationic starch, adsorption equilibrium (or a very slow change in the adsorbed amount) was generally attained within 30-60 min. Table 2 gives the resulting final changes in frequency and damping coefficient for two starches. The frequency changes indicate that, independently of the ionic strength, the adsorbed amount of CS with high molecular weight and low degree of substitution, CSHL, was higher than that of the starch with low molecular weight and high degree of substitution, CSLH. For both starches, the change in frequency increased with increasing ionic strength. Without added electrolyte, the frequency changes are very small and the changes in dissipation are only slightly above the detection limit. Thus, the adsorbed layers are thin and rigidly attached to the surface and it is possible to calculate the adsorbed mass using the Sauerbrey equation (eq 1). The results are given in Table 2. The adsorbed mass is, indeed, very low, corresponding to a layer thickness of 0.3-0.6 nm (assuming that the density of the adsorbed layer is ≈1 g cm-3). Thus, both CSHL and CSLH adsorb on the surface in a very flat conformation. (16) Johansson, L.-S.; Campbell, J. Surf. Interface Anal. 2004, 36, 1018.

Figure 5. Change in the dissipation factor as a function of the change in frequency for adsorption of 100 ppm CS in 100 mM NaCl on a silica surface. Key: ([) low DS (CSHL); (2) high DS (CSLH). f0 ) 5 MHz, n ) 3.

This is not surprising as the repulsion between the ionic groups on the polymer and the polymer/surface attraction are expected to be quite strong at this low ionic strength. On addition of 1 mM simple electrolyte, the changes in frequency and dissipation increase for both starches, but more so for the one with lower DS, CSHL. For CSLH, the frequency change is almost negligible, but the increase in dissipation indicates that the layer has become less rigid. Nevertheless, the dissipation is still so low that it is reasonable to assume that eq 1 can be applied. The results (Table 2) show that the adsorption of CSHL has about doubled, while there is little change in the adsorption of CSLH. The frequency and dissipation changes resulting from adsorption in 100 mM NaCl are substantially higher. Figure 5 shows that the dissipation did not change linearly with frequency, which is a clear indication that the adsorbed layers are soft and less rigidly attached than at low electrolyte concentrations. The approximations underlying the Sauerbrey equation are not valid for such layers. An analysis in terms of the Voigt model will be discussed below. Qualitatively, the results presented in Table 2 are in full accordance with current theory of polyelectrolyte adsorption. CS is a highly water-soluble cationic polymer, and hence, the driving force for adsorption on the hydrophilic, negatively charged silica surface is mainly the attraction between the oppositely charged diffuse double layers. With increasing ionic strength, both this attraction and the repulsion between the charges on the CS are weakened, resulting in less rigid and more weakly bound adsorbed layers. XPS Investigation of the Adsorbed Layers. Two measures of the amount of adsorbed polymer were extracted from XPS spectra of dried adsorbed layers on the QCM crystals: the percentage of nitrogen atoms and the relative

10904

Langmuir, Vol. 20, No. 25, 2004

Tammelin et al. Table 3. Change in Frequency and Dissipation at Different Electrolyte Concentrations for 100 ppm CS Solutions at 15, 25, and 35 MHz CSLH CSHH Mw ) 4.5 × 105 Mw ) 8.7 × 105 overtones [NaCl] (mM) (MHz) ∆ f (Hz) ∆D × 10-6 ∆ f (Hz) ∆D × 10-6 0 100

Figure 6. The amounts of nitrogen and relative abundance of C-O bonds (atom %) in the adsorbed layers of cationic starch analyzed by XPS. The inaccuracy of the measurements was estimated with confidence level 95%, assuming that measurements were normally distributed.

Figure 7. Change in dissipation factor as a function of the change in frequency for adsorption of (a) 10 ppm CSHL (Mw ) 8.8 × 105, DS ) 0.2) solutions and (b) 10 ppm CSLH (Mw ) 4.5 × 105, DS ) 0.75) on a silica surface. Key: ()) CS in 100 mM NaCl; ([) CS in 500 mM NaCl. f0 ) 5 MHz, n ) 3.

abundance of C-O bonds. The latter can be calculated from the high-resolution C 1s XPS spectrum and can be taken as a measure of the amount of carbohydrate in the layer. Figure 6 shows the amounts (atom %) of nitrogen and the relative abundance of C-O bonds in the adsorbed layers of CSLH and CSHL. The nitrogen content of layers adsorbed from solutions without any added simple elec-

15 25 35 15 25 35

-6.0 -9.5 -15.8 -73.7 -94.9 -115.8

0.04 0.00 0.01 4.21 3.50 3.10

-2.5 -4.2 -6.0 -58.3 -81.8 -106.2

0.26 0.01 0.16 4.51 3.74 3.28

trolyte is 0.18% for both polymers, while the amount of C-O is about 3 times higher for CSHL. This is readily explained by the higher nitrogen content of CSLH and indicates that the reaction of the trimethylammonium groups with the negative charge on the surface is probably charge equivalent. The XPS results for adsorption from 100 mM NaCl solutions are also in qualitative agreement with the QCM analysis. Less is adsorbed of CSLH than of CHL, as also indicated by the frequency changes in Table 2. However, the amount of adsorbed starch indicated by the C-O analysis implies that there should be about 3 times as much nitrogen adsorbed with CSLH as with CSHL. The ratio indicated by the XPS analysis is only about 2. Possible reasons for this will be discussed below. Adsorption at Higher Electrolyte Concentrations. Figure 7 shows ∆D versus ∆f for adsorption of CSHL and CSLH at higher electrolyte concentrations. Note that the initial concentrations of the two starches are only 1/10 of those in Figure 5. Despite the much lower concentrations, the final values of ∆f are about the same as for 100 ppm CS and also virtually independent of the ionic strength, indicating that the plateau level of the adsorption isotherm is reached already in 10 ppm CS solutions and that this level is not much changed by adding more electrolyte above 100 mM. For CSHL, increasing the electrolyte concentration to 500 mM results in a considerably increased dissipation, in particular in the saturated layer, while the trend is the opposite for CSLH. A qualitative interpretation of the difference between the two polymers would be the following. The layer attached to the silica surface and sheared by the oscillating crystal contains both adsorbed polymer and water. The dissipation mechanisms will depend on the rigidity of the polymer and the amount of water that is sheared together with the polymer. The polymer with low DS is loosely bound to the surface and becomes more so as the electrolyte concentration increases. This would result in more mobile polymer chains and a larger amount of water that moves with the polymer. Hence, dissipation increases as the electrolyte concentration increases. The polymer with high DS is strongly bound to the surface, and the main effect of adding salt is that repulsion between the polymer segments is screened so that the polymer coils contract somewhat and expel water. Hence, less water moves with the sheared layer and the dissipation decreases. A more detailed analysis of the layer properties will be given below. The Effect of Molecular Weight and Molecular Weight Distribution. The results shown in Figure 7 can also be discussed in terms of the molecular weight effect on the adsorption. Figure 1 shows that CSHL (Mw ) 8.8 × 105, DS ) 0.2) and CSLH (Mw ) 4.5 × 105, DS ) 0.75) have narrow molecular weight distributions. It can be assumed that electrostatic interactions are weak when the NaCl

Viscoelastic Properties of Cationic Starch

Langmuir, Vol. 20, No. 25, 2004 10905

Figure 8. Change in dissipation factor as a function of the change in frequency for adsorption of 10 ppm CS (DS ) 0.75) in 100 mM NaCl. Key: ([) Mw ) 8.7 × 105; (9) Mw ) 4.5 × 105. f0 ) 5 MHz, n ) 3.

Figure 10. The change in frequency as a function of time for CSHL (Mw ) 8.8 × 105, DS ) 0.2). f0 ) 5 MHz, n ) 3.

Figure 9. The amounts of nitrogen and relative abundance of C-O (atom %) of the adsorbed layers of cationic starch with DS ) 0.75 analyzed by XPS. The NaCl concentration was 100 mM. The standard deviations were estimated with the confidence level of 95% assuming that measurements were normally distributed. Table 4. Effect of the Solution Concentration on the Adsorbed Amount of Cationic Starcha CS concentration CSHL CSLH CSHH a

∆f (Hz) ∆D × 10-6 ∆f (Hz) ∆D × 10-6 ∆f (Hz) ∆D‚10-6

10 ppm

100 ppm

-32 0.49 -8.7 0.03 -6.4 0.29

-10.5 0.14 -6.0 0.04 -2.5 0.26

f0 ) 5 MHz, n ) 3.

concentration exceeds 100 mM, so that differences in the adsorption behavior of these two polymers at high salt concentrations will be due to effects of the molecular weight. The ∆D/∆f value at the end of the adsorption for both polymers under these conditions is ∼36 (see curves Ia and IIb in Figure 7), which indicates layers with similar structure and properties. The final change in frequency for CSHL in 100 mM NaCl is ca. -100 Hz, whereas the frequency change for the CSLH in 500 mM NaCl is ca. -70 Hz. Thus, the adsorbed amount increases with increasing molecular weight but other properties of the adsorbed layer are not significantly affected by the size of the molecule. Figure 1 shows that while CSHH and CSLH both have a peak in the molecular weight distribution at Mw around 5.5-6 × 105, CSHH has a very broad tail toward high molecular weight. Table 3 shows the amounts adsorbed of these two starches at different ionic strengths. Without any added electrolyte, very small amounts of either polymer are adsorbed and the changes in dissipation

are negligible. Indeed, the frequency changes at 15 MHz (2nd overtone) are almost negligible in comparison with the noise and drift of the instrument, but the results at higher overtones show that adsorption does, indeed, take place. Without added salt, both polymers obviously are adsorbed in a very flat conformation. At higher salt concentrations, larger amounts are adsorbed, but the frequency and dissipation changes both indicate that larger amounts of the low molecular weight polymer CSLH adsorb in a more compact conformation than CSHH (Figure 8). That the amount of CSLH is larger than the amount of CSHH is confirmed by XPS analysis (Figure 9), which shows that both the nitrogen content and the amount of C-O in the dried layer are higher for CSLH. Figure 9 also shows that the amounts adsorbed from 10 and 100 ppm CS solutions are the same; that is, the adsorbed layer is saturated already at 10 ppm. For polydisperse polymers, one expects the molecular weight distribution on the surface to be shifted toward higher molecular weights than the distribution in solution.1 Thus, it can be concluded from the distributions in Figure 1 that the high molecular weight tail of CSHH should be preferentially adsorbed and, thus, that the difference in molecular weight distributions of CSLH and CSHH on the surface will be even more marked than in solution. The QCM results confirm this prediction. While at low salt concentration both polymers are tightly bound, the low molecular weight CSLH is able to adsorb in a closely packed, dense layer while CSHH is not able to pack as closely and, hence, a lower total mass of polymer is adsorbed. On adding salt, both polymers become more coiled and dissipation increases. This effect should be more marked for the polymer with high molecular weight, and accordingly, dissipation increases more rapidly and the final amount adsorbed is smaller for CSHH (Figure 8). The Effect of Concentration. During diffusioncontrolled adsorption, smaller molecules will reach the surface more rapidly than larger ones. This should further enhance the difference between CSLH and CSHH in the direction described above; at a given concentration, a densely packed layer of small molecules is rapidly formed by CSLH, while the CSHH layer contains larger and more loosely packed molecules, see Table 4. The importance of this effect is verified by the adsorption behavior of CSHL shown in Figure 10. For this polymer and also for the two other polymers, the initial rate of adsorption is proportional to the polymer

10906

Langmuir, Vol. 20, No. 25, 2004

Tammelin et al.

Figure 12. (a) Shear viscosity and (b) shear modulus as a function of time for adsorption of CS from 100 mM NaCl. The initial concentration of CS is 10 ppm. Key: ([) CSHL (Mw ) 8.8 × 105, DS ) 0.2); (thin line) CSLH (Mw ) 4.5 × 105, DS ) 0.75); (thick line) CSHH (Mw ) 8.7 × 105, DS ) 0.75).

Figure 11. Adsorption of CS from 10 ppm solutions in 100 mM NaCl. ∆f and ∆D versus time for n ) 3, n ) 5, and n ) 7 (lines indicate QCM-D data) and the best fit obtained using the Voigt model (squares indicate fitted values). (a) CSHL (Mw ) 8.8 × 105, DS ) 0.2); (b) CSLH (Mw ) 4.5 × 105, DS ) 0.75); (c) CSHH (Mw ) 8.7 × 105, DS ) 0.75). Note that only in the case of the lowest curves in panels b and c is the difference between calculated and experimental values really discernible (lower curve calculated, upper curve experimental).

concentration, which is typical for a diffusion-limited process. The final adsorbed amount increases substantially when the concentration is increased from 1 to 10 ppm but decreases again when the solution concentration is increased to 100 ppm. The concomitant change in dissipation is close to zero for adsorption from 1 and 100 ppm solution concentrations, while the layer adsorbed from a 10 ppm solution gives a dissipation change of 0.5 × 10-6. A tentative explanation for this behavior would be the following: At very low concentrations, adsorption is slow and there is time for the starch molecules to settle down on the surface. In 10 ppm solutions, both small and large molecules are adsorbed in a more coiled conformation,

leading to higher adsorption, but the layer is still fairly rigid as indicated by the low dissipation. In 100 ppm solution, the surface is immediately covered by rapidly diffusing small molecules that prevent further adsorption. Figure 10 compares adsorption from solutions without added salt. At constant and low ionic strengths, the behavior is essentially the same as with salt, but in 100 ppm NaCl the final amount is large and no longer depends on the total concentration of CS (as for CSHL and CSHH in Figure 9) because repulsion between the adsorbed molecules is low, attraction between the surface and the polymers is lowered, and the molecules are adsorbed in a coiled conformation. Discussion Viscoelastic Properties of the CS Layers. In the fitting of the results to eqs 4-7, it is assumed that the density Fl and viscosity ηl of the bulk liquid are those of water. The effects of dissolved starch and salt on the values of Fl and ηl are very small and, hence, were neglected. In the modeling of the polymer film properties, the density of the film Ff was assumed to be constant (1.2 g cm3) and the thickness hf, viscosity ηf, and elasticity µf were adjusted to give the best possible fit using a least-squares method. Obviously, describing the complex relaxation mechanism and different contributions to the elastic properties of an adsorbed polymer layer with a Voigt element is a gross

Viscoelastic Properties of Cationic Starch

Langmuir, Vol. 20, No. 25, 2004 10907

Table 5. Final Shear Moduli and Viscosities as Well as Relaxation Rates in the Beginning and at the End of Adsorption of Cationic Starches from 100 mM NaCl ηf × 103 (N s m-2)

µf × 105 (N m-2)

τf (ns)

sample

20 min

90 min

20 min

90 min

20 min

25 min

90 min

CSHL (Mw ) 8.8 × 105, DS ) 0.2) CSLH (Mw ) 4.5 × 105, DS ) 0.75) CSHH (Mw ) 8.7 × 105, DS ) 0.75)

2.08 1.9 1.2

2.12 1.8 1.1

1.6 1.6 1.3

2.1 1.4 1.1

14 12 9.2

11

9.7 13 10

oversimplification. However, despite its simplicity, the Voigt model can be fitted to the frequency and dissipation curves surprisingly well. Introduction of additional features which necessarily must also involve more fitting parameters would be reasonable only if more overtones and less noisy QCM measurements are used. For lack of such data, application of a more complex model than the Voigt element does not seem to be warranted. The simplicity of the model implies that while the results can be compared relative to each other, not too much significance should be attached to the absolute values of the viscosities and elasticities obtained. Adsorption from Solutions without Added Simple Electrolyte (NaCl). The results show that layers formed by adsorption of CS without any added simple electrolyte are thin and rigidly attached onto the anionic surface. The change in dissipation is less than 1 × 10-6. This is too low for meaningful interpretation in terms of viscoelastic modeling. The layers are fully elastic, and the thickness of the films can be estimated by using the Sauerbrey eq 1. Adsorption from 100 mM NaCl. At high ionic strength (100 mM NaCl), the viscoelastic properties and the final thickness of the cationic starch layer can be estimated using the Voigt based model. Figure 11a-c shows the measured and fitted ∆f and ∆D curves for adsorption from 10 ppm solutions of the three cationic starches in 100 mM NaCl. Figure 12 shows the shear viscosities and shear moduli corresponding to the fitted curves in Figure 11. Relaxation times (ηf/µf) after 20 min and at the end of adsorption are given in Table 5. Finally, Figure 13 shows the hydrodynamic thickness if the layers are calculated by the application of the Voigt model. The results show that, despite the screening of the electrostatic interactions by the simple electrolyte, the layers still behave as expected for adsorbing polyelectrolytes. The attraction of the CS with low charge density, CSHL, to the surface is weak, and the polymer forms a thick and mobile layer (short relaxation time) with higher elasticity and viscosity than the layers formed by CSLH or CSHH. As the layer thickness increases, the relaxation time decreases: more and more large molecules weakly bound to the surface are adsorbed. The starches with higher charge density are more strongly bound to the surface. The layers are considerably thinner than for CSHL, and as adsorption proceeds they become more tightly packed and less mobile: the relaxation times increase. The thicknesses of the layers of CSLH and CSHH are about the same, but the viscosity and elasticity of CSLH are higher. As already noted in the description of the primary ∆f/∆D data, this polymer is probably more densely packed than CHH. Adsorption from 500 mM NaCl. In 500 mM NaCl, electrostatic interactions should be effectively screened and the starch is expected to behave more or less like a neutral polymer. It is interesting to note that both CSHL and CSLH are still adsorbed. The fitting of the Voigt model to the frequency and dissipation data still is quite good, except for the frequency at the 7th overtone (Figure 14a).

Figure 13. The hydrodynamic thickness as a function of time (the adsorbed amount) in 100 mM NaCl. Key: ([) CSHL (Mw ) 8.8 × 105, DS ) 0.2); (thin line) CSLH (Mw ) 4.5 × 105, DS ) 0.75); (thick line) CSHH (Mw ) 8.7 × 105, DS ) 0.75).

Viscosities and elasticities are given in Figure 15, and some relaxation times are given in Table 6. Figure 16 shows the hydrodynamic thickness if the layers are calculated by the application of the Voigt model. The most remarkable difference from the behavior in 100 ppm NaCl is that the viscosities and elasticities of CSHL and CSLH now depend in the same way on adsorption time and the properties of both layers are now remarkably similar. This is not surprising, as all effects of charge density have been swamped out. The layer thicknesses are about the same as in 100 ppm NaCl and, as expected, larger for the polymer with high molecular weight. The viscosities are slightly higher than in 100 ppm NaCl, while the elasticities of both polymers are the same as that of CSLH in 100 ppm NaCl. Accordingly, the relaxation times are shorter: removal of all electrostatic interactions between chain segments has led to a more mobile layer. As expected, the relaxation times for both layers increase slightly as the layer becomes more densely packed. Adsorption at High Electrolyte Concentrations. The theory of polyelectrolyte adsorption1 predicts that the maximum adsorbed amount of a polyelectrolyte should be lower, the higher its charge density, because less polymer is required for charge neutralization and the adsorbed layer is flatter. Addition of simple electrolyte screens intramolecular repulsion between polymer segments as well as polymer/surface interactions. Both effects lead to an increase in the maximum amount of adsorbed polymer because the persistence length of the polymer decreases so that its conformation on the surface is less extended and the polymer forms a thicker layer until finally polymer/surface interactions become so weak that the polymer does not adsorb at all (unless there are nonelectrostatic interactions that promote adsorption). In dilute solutions, polyelectrolyte molecules expand strongly due to the osmotic pressure created by the counterions. The expansion of polyelectrolytes decreases with decreasing polymer charge density and increasing ionic strength. Fewer salt ions are needed to repeal the

10908

Langmuir, Vol. 20, No. 25, 2004

Tammelin et al.

Figure 14. ∆f and ∆D versus time at n ) 3, n ) 5, and n ) 7 (lines indicate QCM-D data) and the best fit obtained using the Voigt model (squares indicate fitted values) for adsorption of CS from 500 mM NaCl. The initial concentration of CS is 10 ppm. (a) CSHL (Mw ) 8.8 × 105, DS ) 0.2); (b) CSLH (Mw ) 4.5 × 105, DS ) 0.75). Table 6. Shear Moduli, Viscosities, and Relaxation Rates in the Beginning and at the End of Adsorption of Cationic Starches from 500 mM NaCl ηf × 103 (N s m-2) sample CSHL (Mw ) 8.8 × 105, DS ) 0.2) CSLH (Mw ) 4.5 × 105, DS ) 0.75)

Figure 15. Variations in (a) shear viscosity and (b) shear modulus as a function of time. Key: (thin line) CSHL (Mw ) 8.8 × 105, DS ) 0.2); (thick line) CSLH (Mw ) 4.5 × 105, DS ) 0.75).

osmotic pressure in the case of low charged polymer compared to highly charged polymer.17,18 As shown by Table 2, the cationic starch behaves exactly in this way. Without added electrolyte, starch with high charge density adsorbs on the anionic silica in a very flat conformation and the amount detected is very small. (17) Gupta, P. R.; McCarthy, J. L. Macromolecules 1968, 3, 236. (18) Mabire, F.; Audebert, R.; Quivoron, C. Polymer 1984, 25, 1317.

µf × 10-5 (N m-2)

τf (ns)

20 min 90 min 20 min 90 min 20 min 90 min 2.0

1.6

2.8

2.0

7

8

1.6

1.5

1.9

1.4

8

11

Addition of simple electrolyte reduces the persistence length and reduces polymer/surfactant interactions so that more polymer fits on the surface and adsorption rises. The conformation of starch with low charge density is sufficiently coiled already when no salt is added, so that adsorption without added NaCl is already significant. The adsorbed amount increases with increasing concentration of NaCl. Application of eq 1 yields adsorbed amounts that are within the typical range expected for relatively high molecular weight polymers (1-3 mg/m2). As shown by the ∆D/∆f plot (Figure 5), the adsorbed layer in 100 mM NaCl is very loose so that dissipation rises steeply; the calculation of adsorbed mass from the Sauerbrey equation then becomes unreliable. The XPS analysis supports the QCM-D results (Figure 6). More starch is adsorbed on the silica surface when starch with a low degree of substitution is used or the electrolyte concentration is increased. Starch film also seems to be uniformly attached onto the surface; the standard deviation of the N and C-O percentages of the surfaces is fairly low. Further increase in electrolyte concentration (500 mM NaCl) shows the effect of increased amount of counterions together with the polymer charge on the conformation of the molecule and on the adsorption/desorption behavior. A layer of CS with low charge density has a high water content and binds very loosely to the surface (dissipation rises very steeply with ∆f, ηf and µf decrease, τf decreases), whereas highly charged CS forms a more and more coiled polymer layer (dissipation increases less steeply with ∆f, ηf and µf decrease, τf increases) (Figure 7). Table 7 summarizes the effects of polyelectrolyte charge density at high ionic strength on the adsorption of CS and on the viscoelastic properties of the formed layer. Adsorption from 500 mM NaCl makes it possible to compare the effect of molecular weight without electro-

Viscoelastic Properties of Cationic Starch

Langmuir, Vol. 20, No. 25, 2004 10909

Table 7. Summary of the Effects of Cationic Starch Charge and Electrolyte Concentration on Adsorption and Viscoelastic Properties of the Formed Film CSHL Mw ) 8.8 × 105, DS ) 0.2, narrow Mw distribution 100 mM NaCl ∆D/∆f (10 ppm) C-O (%) N (%) ηf (N s m-2) µf (N m-2) τf (ns) hf (nm)

36 12.2 ( 0.2 0.46 ( 0.04 (1.8 ( 0.3) × 10-3 (23 ( 5) × 104 7.8 6.9 ( 0.2

CSLH Mw ) 4.5 × 105, DS ) 0.75, narrow Mw distribution

500 mM NaCl 56 1.6 × 10-3 14 × 104 11 7.0

100 mM NaCl 48 9.8 ( 0.9 0.86 ( 0.09 (1.6 ( 0.1) × 10-3 (16 ( 3) × 104 10 5.1 ( 0.1

500 mM NaCl 34 1.3 × 10-3 20 × 104 6.5 4.1

Table 8. Summary of the Effects of Molecular Weight Distribution of Cationic Starch on Adsorption and Viscoelastic Properties of the Formed Film CSLH, Mw ) 4.5 × 105, CSHH, Mw ) 8.8 × 105, DS ) 0.75, DS ) 0.75, narrow distribution broad distribution

Figure 16. The hydrodynamic thickness as a function of time (the adsorbed amount) in 500 mM NaCl. Key: (thin line) CSHL (DS ) 0.75, Mw ) 4.5 × 105); (thick line) CSLH (DS ) 0.75, Mw ) 8.7 × 105).

static contributions. Adsorption of polyelectrolytes on smooth surfaces increases with increasing molecular weight. Comparison of the adsorption of CSHL in 100 mM NaCl with the adsorption of CSLH in 500 mM NaCl (Table 7 and Figure 7) shows that the polymers follow this rule. Increasing molecular weight leads to a larger hydrodynamic thickness. In diffusion-controlled adsorption, the smaller molecules migrate to the surface faster than larger ones. Comparison of CSHH and CSLH shows that less adsorbs of the high molecular weight CS with a broad distribution than of the low molecular weight CS with a narrow distribution (Table 3 and Figure 8). XPS analysis supports this conclusion (Figure 9). Table 8 summarizes the effect of molecular weight and molecular weight distribution on the adsorption of CS and on the viscoelastic properties of the formed layer. Figure 17 schematically illustrates the structure of the adsorbed layers. It is clearly seen from Table 8 that CSLH which has a narrow distribution of molecular weights forms a more compact and rigid layer (lower ∆D/∆f, higher ηf and µf) due to the better packing of the starch molecules than the broadly distributed CSHH. The hydrodynamic thicknesses of the two layers do not differ much. Concluding Remarks Previously, we have shown that cationic starch in solution retains the coiled conformation of native starch despite the rather drastic changes introduced by the cationic substituents.19 Nevertheless, on adsorption, the dependence of the structure and viscoelastic properties of the adsorbed layer on charge density, molecular weight, (19) Merta, J.; Garamus, V. M.; Kuklin, A. J, Stenius P.; Willumeit, R. Langmuir 2000, 16, 10061.

∆D/∆f (10 ppm solution) C-O (%) N (%) ηf (N s m-2) µf (N m-2) τf (ns)a hf (nm)

48.0

72.3

9.8 ( 0.9 0.86 ( 0.09 (1.6 ( 0.1) × 10-3 (160 ( 30) × 103 9.2/10 5.1 ( 0.1

7.0 ( 1.6 0.6 ( 0.2 (1.10 ( 0.03) × 10-3 (94 ( 6) × 103 12/13 4.8 ( 0.4

a Relaxation times calculated for adsorption at 20 min/90 min, see Table 4.

Figure 17. Schematic drawing of the structure of the layer of (a) CS with narrow molecular weight distribution and (b) CS with broad molecular weight distribution. v1 and v2 represent the diffusion rates of the molecules of different sizes (v1 > v2).

and electrolyte concentration does not differ significantly from the well-known behavior of randomly coiled linear polycations. The highly simplified Voigt element-based rheological model describes the viscoelastic properties of the adsorbed layers surprisingly well. An observation that merits further investigation is that not even the starch with a lower degree of substitution (0.2) is desorbed at high electrolyte concentrations. Apparently van der Waals interactions are sufficiently strong to keep the molecules attached even in the absence of significant electrostatic forces. Acid/base interactions between the acidic silanol groups on hydrated silica and the basic trimethylammonium groups may also play a role. Acknowledgment. This research was supported by a grant from the Finnish Technology Agency, TEKES. LA0487693